[Univ of Cambridge]alt[Dept of Engineering]

MIL Speech Seminars 2005-2006

The MIL Speech Seminar series schedule for the Long Vacation 2006 was as follows:

10th July 2006 Professor Stephen Cox (UEA) Automatic Determination of Musical Genre and Similarity Recently, there has been a revolution in the way that music has been delivered to users. The universal availability of broadband to the home and the development of cheap, high-capacity MP3 players has led to an exponential growth in music distribution over the internet, and to the emergence of large personal collections of "songs" held on users' computers and players. This in turn has led to a need for effective techniques for organising, browsing and visualising music collections and generating playlists. Although metadata giving details of e.g. the track title, the album, the artists etc. is available for much of the music available on the web, it is not universal, and this data is usually not detailed enough to implement the above techniques to a high standard. We have been investigating techniques for automatically classifying the genre of a song and measuring the similarity of two songs using only the audio signal. I will describe our approach to these two related tasks, and present results that suggest it is possible to perform them with reasonable accuracy. I will also demonstrate our playlist generator that suggests songs similar to an input song from a 5000 song collection.
2nd August 2006 Dr Paul Taylor (MIL Lab) Progress in Speech Synthesis This talk is to give an overview of what I have been doing in my two years at the MI lab. It will cover a number of related topics including (i) a new model of prosodic semantics and function, and how this affects speech technology applications, (ii) a statistical model of text analysis based on the source/channel decoing paradigm, (iii) a new grapheme to phoneme conversion technique based on HMMs, (iv) an overview of the new speech synthesizer I have developed which (a) unifies HMM and unit selection synthesis, (b) allows a number of target and join cost formulations to be used, (c) has new way of thinking about target costs and decision trees in synthesis, (d) has new statistical target and join cost algorithms, (v) A quick overview of the book I have been working on.